Recent advances in heart sound analysis
[EN] Objective: Auscultation of heart sound recordings or the phonocardiogram (PCG) has been shown to be valuable for the detection of disease and pathologies (Leatham 1975, Raghu et al 2015). The automated classification of pathology in heart sounds has been studied for over 50 years. Typical metho...
| Autores: | , , , , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2017 |
| País: | España |
| Institución: | Universitat Politècnica de València (UPV) |
| Repositorio: | RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia |
| Idioma: | inglés |
| OAI Identifier: | oai:riunet.upv.es:10251/152717 |
| Acceso en línea: | https://riunet.upv.es/handle/10251/152717 |
| Access Level: | acceso abierto |
| Palabra clave: | Heart sound Signal processing Physiological Measurement Adquisition Detection TECNOLOGIA ELECTRONICA |
| Sumario: | [EN] Objective: Auscultation of heart sound recordings or the phonocardiogram (PCG) has been shown to be valuable for the detection of disease and pathologies (Leatham 1975, Raghu et al 2015). The automated classification of pathology in heart sounds has been studied for over 50 years. Typical methods can be grouped into: artificial neural network-based approaches (Uguz 2012), support vector machines (Ari et al 2010), hidden Markov model-based approaches (Saracoglu 2012) and clustering-based approaches (Quiceno-Manrique et al 2010). However, accurate automated classification still remains a significant challenge due to the lack of highquality, rigorously validated, and standardized open databases of heart sound recordings. Approach: The 2016 PhysioNet/Computing in Cardiology (CinC) Challenge sought to create a large database to facilitate this, by assembling recordings from multiple research groups across the world, acquired in different real-world clinical and nonclinical environments (such as in-home visits), to encourage the development of algorithms to accurately identify, from a single short recording (10-60s), as normal, abnormal or poor signal quality, and thus to further identify whether the subject of the recording should be referred on for an expert diagnosis (Liu et al 2016). Until this Challenge, no significant open-access heart sound database was available for researchers to train and evaluate the automated diagnostics algorithms upon (Clifford et al 2016). Moreover, no open source heart sound segmentation and classification algorithms were available. The Challenge changed this situation significantly. Main results and Significance: This editorial reviews the follow-up research generated as a result of the Challenge, published in the concurrent special issue of Physiological Measurement. Additionally we make some recommendations for promising research avenues in the field of heart sound signal processing and classification as a result of the Challenge. |
|---|